All too often game structures are designed to extrinsically motivate the player solely for the sake of retention. Games desperately hope that players will find some intrinsic enjoyment while they're being tricked into wasting their time. Or worse yet, the retention is itself the point and then games hope to create compulsive behavior in the player with gambling-like mechanics. Devs treat players like morons, and even if it's correct, it ain't right.
I think this completely inverts the relationship intrinsic vs extrinsic motivation mechanics should have, especially as it pertains to arcadey or challenge-based games of all types. It's bad artistically because it attempts to homogenize how games are structured. It's bad practically because it creates surface-level engagement, where people play for the structure, not for the underlying game. These cheap tricks won't work for too long in games that are naturally discouraging & difficult.
Engagement with these games is deeply intrinsically driven at its core - players struggle against these games because they want to improve. Something about the process is fulfilling to them on a personal level. The game structures should not be tricks which attempt to fool the player to engage, they should be tools that help players voluntarily structure their engagement.
Scoring is perhaps the most obvious example. It can be confused with other extrinsic motivators like idk, unlocking hats, but it's very different under the surface.
It means absolutely nothing until you opt into its risk vs reward structures, which there's often no reason to do. Many games let you hit extra life caps without paying attention to score. Oftentimes the games don't even telegraph scoring via kinaesthetic feedback (like coins, medals, cubes, big flashing multipliers), so scoring is even more voluntary than doing things solely because they feel cool or good on a visceral level.
Even outside of scoring, all you need to do in order to understand the role of structures in arcadey/challenge-based games is look at what players do in games with no such structures, such as most arcade games or the more barebones ports.
As players start getting into a game, they will start building their own incentive and reward structures. They will come up with short and long term goals, they will try to look at their smaller improvements, they will do competitions, they will talk about the games with others.
The players also create reward structures - they upload replays, they annotate them, they write guides, they share them with others, they do commentary videos, they even do analysis. All of that gives more meaning to their improvement, it's them "cashing out" so to speak. The extrinsic structures are being used to enhance their intrinsic desire to play & improve.
These behaviors give a good guide for developers who want to use extrinsic motivators/structures in a healthy way. Give players tools to alleviate frustration and keep themselves going when they stumble. Assume that players already want to play with your game because it's fun, don't try to trick players into engaging.
With that in mind, here are some concrete examples of what I'd consider healthy game structures for a genre like shmups :
- A variety of sub-modes which get players to engage with a different side of the game they're playing, while still feeding into their general skillset. Racing games do this very well - the core skills being tested are your intuitive understanding of the physics engine and your ability to come up with good lines or strats. Different modes like hotlap, battle modes, drifting modes, drag modes, etc. might hyper focus on specific mechanics, but it all feeds back into going fast in one way or another. Shmups could even have "routing puzzle" modes which would test the player's knowledge of the scoring system.
- Sub-modes create meaningful specialization. A player who's good at coming up with strats might not be good at playing overall, but if there's a mode for them then they can specialize & give back to other players. You see this happen a lot in speedrunning.
- Sub-goals within sub-modes. Practice can feel aimless, so creating small goals & self imposed challenges within practice modes can help alleviate that feeling of aimlessness. Maybe have some goals for timing out the boss, maybe have some goals for quick killing, some goals for no bombing, some goals that reward bomb routing. There can even be meta-goals like stats tracking how many attempts are successful, so players can see increases in consistency.
- Convenient, robust replay systems. Replays are the player's pride and joy. They are a concrete tangible result of their strugges, like a painter finally finishing a piece instead of stopping at sketches. Letting players own the replay and make it their own is vital. Let them commentate their replay, let them customize parts of it by changing colors, music, extra visual elements etc. Let them make their replay theirs.
- Graphs and stats showing players their improvement over time. It's very important to capture their increased consistency and shove it in their face a bit. Otherwise they don't realize that they are learning.
- Fun, convenient competitive modes. Either Z-Chaser esque competition between the player and their leaderboard neighbour, or full blown indirect PvP modes similar to the kumite modes that have gotten popular at shmup events. It gives something concrete to work towards, and shifts the player's focus away from single goals towards consistency.
It's a tricky balance to strike because even "healthy" structures can become addictive. But the general idea behind all of this should be to try and stay in line with what the games are at their core - difficult challenges that reward players who step up. They aren't brainless, passive, flow optimized experiences that benefit from endless unlockables, but if devs treat them like they are then eventually that's what they will become.